Fri 2:40 PM

My adventures begin. I am packed, sitting in my room, waiting for Chris Brown Denny to come pick me up. He has not responded to my messages. I have already missed my alternative way to Houston. CBD where are you.

Fri 6:47 PM

After about 3 hours or so of driving Chris and I have made it. Traffic was ok, Chris drives pretty fast, I was definitely scared at some points. But we made it. Found the bus over to the event, it’s right next to the Brochstein Pavilion, which is really exciting! Got checked in, got a shirt, found Eric (my partner), now they’re going through the possible prizes. Zzzzzz

Fri 8:01 PM

Eric and I have just gotten dinner. I’m very impressed. It’s Thai. It’s very good. We are also coming to settle on an idea. Something something future prediction bitcoin arbitrage…

Fri 8:34 PM

Crisis. Our idea has holes. Future arbitrage…. is actually just trading. Everyone already does digital coin trading! We are ruined.

Fri 9:11 PM

Crisis resolved. Realized that it’s a hackathon so semantics don’t matter.

Fir 11:06 PM

Eric went to Ozzies. He has abandoned me on my own here. Fortunately, I am a hackathon veteran. I know how to survive on my own.

The wifi is failing me. I need to migrate to higher ground. I scrounge up all useful materials around me; the leftover oreos, all the bbq and sour cream lays in the chips box, an abandoned unopened water bottle, and start trying to find places where wifi will actually connect. Tiffs treats arrive. I get one, then go through the line again, and get two and stash two more away. My wifi troubles are not alleviated. Perhaps restarting my computer will help.

Fri 11:11 PM

A restart fixed it. I suspect that was all I needed from the beginning. I like my new vantage point more anyways. The recruiters have gone home, and other hackers are already beginning to raid their booths for goodies.

Sat 1:02 AM

It seems that Eric did not go home merely to take a shower. He is going to bed. Another couple has sat down at my table, which used to be empty, and are friendly. Overly friendly. They introduce themselves and ask about my project, and discuss theirs. Loudly. My focus is shattered. And to top it off the wifi starts dropping again….

Sat 1:47 AM

Ryan Rice materializes. Maybe he will have answers. I explain our ideas to him. He looks like he does not care. He looks like this was more than he bargained for. I need answers. He has none. He is here to eat cookies and play smash. He wriggles away. I get desperate enough for hot water/tea that I set a coffee machine going without any coffee grounds. They don’t have teabags. I guess I’ll be having hot milky water.

Sat 2:24 AM

Yes, I am trying to find negative cycles using a bellman ford algorithm. Negative cycles of negative prices. I don’t want to find cycles that actually make me a net negative money. Why is this happening.

Sat 8:56 AM

I am awake. I gave up and went to bed (read: wrapped a blanket around myself on the hard floor) around 3AM last night in a state of despair. I wake up to my phone’s alarm. I didn’t set an alarm. Oh yeah, I set one when fooling around with NSC chips. My mood has improved. I think I have a way this will still work.

Sat 1:02 PM

Lunch was potbelly sandwiches. Barely acceptable. A letdown after the Thai of last night. I pilfer three anyways. I’m hard at work changing my mind on how I want to store the data for the 3rd time. Eric and I have come to the conclusion that our problem is some variant on a standard search problem, and that we’re too dumb to formulate it correctly in order to find the optimal solution. We decide to just BFS and fill in the whole tree and keep track of the best solutions seen. Thanks AI.

Sat 3:51 PM

Almost crisis. I’ve been installing so many python3 modules that my CS disk quota runs out. I have to go through and delete the dumb .cache and .config files that Chrome creates in my home directory to free up enough space to install fbprophet. Delayed by 15 minutes.

Sat 4:04 PM

There is a “no light” challenge. I am intrigued. The idea is you have to code a webpage without any resources, no google, no syntax highlighting, nothing. I am entirely reliant on google for CSS. Time to destroy these fools.

Sat 4:11 PM

It’s been 10 minutes and they’re teaching me how classes and IDs work… The lady in charge finally stops the presenter and says “hey get on with it.”

Sat 4:14 PM

Finally getting started. Time to CRUSH THESE FOOLS.

Sat 4:29 PM

That was fun. I have 2.5kb of code submitted. Nobody else has even a kb.

Sat 4:37 PM

The flexbox didn’t work, it destroyed the <hr/>s, my fonts weren’t big enough, what a disaster. Still top 50% tho. At least in my opinion.

Sat 4:42 PM

Disaster. While I was wasting time with that insolent competition, an earlier change I made broke data collection. Instead of the 40+ data points we should have, we have 4.

Sat 5:43 PM

I’ve made good progress. Written a pretty gnarly list comprehension (successors += [ (exchange, currency, time + 1, amount) for exchange, data in etd.items() if any([ from_ == currency for (from_, to), _ in data.items() ]) ]). My partner, however, left to ‘chill out’ about an hour ago. He has yet to resurface.

Sat 6:08 PM

This is incredibly strange. I have a function I’m importing. I have print statements in the function, but they’re not being called. I can see them. When I Ctrl+C in IPython it shows it stopping right after the print statement. But it’s not printing.

Sat 6:49 PM

Eric finally shows his face. After I’ve already coded the past data prediction (we make 4 cents). After dinner has been served. After he’s taken a shower. Welcome back Eric.

Sat 7:26 PM

Disaster! I go back for more tacos, but they’ve already been taken away. I only pilfered away three 😭 Eric, a generous soul, donates his last one to me. All is forgiven.

Sat 8:34 PM

Eric writes a function that prints a bunch of output to the console, but returns nothing. He declares it done. “But I need the data elsewere…” “Just look at the output on the console.”

Sat 8:59 PM

We test a run with depth five of the whole system together. We have htop open in another tmux window. We watch as the memory climbs… and climbs… and climbs… to use the entire 16 gigs of ram that the computer has available. Eric wants to kill it, but I’m curious to see if it’ll ever end. So we leave it running and ssh into another lab machine to start an instance with a smaller depth. That one’s memory begins to climb….

Sat 11:38 PM

Things have been coming along at a pretty solid clip. My CLI interface is pretty, Eric is getting work done, looks like we’ll have a pretty neat final project!

Sun 12:12 AM

We’re approaching my favorite part of a hackathon. The part where you can walk around and see the stress and despair on people’s faces. I munch happily on my freshly delivered pizza. I ask Eric what he thinks of his first hackathon. He says that his opinion will primarily be determined by if we win anything or not.

Sun 12:47 AM

I am subtly trying to get Eric hyped / give him small little tasks to keep him going longer and longer - he seems like he’s petering out. I’m ready for that 4 AM high.

Sun 1:06 AM

I am beginning to suspect that our main algorithm has a bug, but I keep my suspicions to myself….

Sun 1:43 AM

I have filled my screen up with graphs. I feel like a real data scientist.

Sun 2:14 AM

The sleep deprevation is hitting hard. For a moment I thought that the jelly scrolling on my one plus five was contagious and had spread to my laptop. The screen was acting wonky! Then I realized it was just my eyes.

Sun 2:22 AM

We’ve decided on a final couple of tasks before we go to bed. I’m starting to feel it too. As evidenced by what I wrote previously I guess ^^^

Sun 2:56 AM

We finally create the github and decide that we’re pretty much done for the night. I say goodnight to Rachel. Time to catch the zzzs I can.

Sun 8:43 AM

I am up. I had some weird dreams. Submissions are due in thirteen minutes. They repeatedly remind us of this through the microphone. And on slack. I think I pulled something in my back and am in pretty major pain if I sit or stand wrong. Less the twenty groups seem to remain on the floor. More will materialize out of thin air when presentation time comes. They need to present to get their travel reimbursement.

Sun 9:06 AM

It’s past nine, which means devposts have been submitted, and hacking is “done”. I am working on modeling exchange fees better.

Sun 9:11 AM

I should keep working, but there are kolaches in the food room.

Sun 9:41 AM

Judging will start in fifteen minutes. I think I found a quick fix to improve accuracy…

Sun 10:00 AM

We’ve found our table. I’m frantically installing the project on my local computer. Here we go.

Sun 10:49 AM

Presentations are underway. I’m getting my usual presentation rush. I’m not getting “winning” vibes from any of the judges though. I sent Eric out on a scouting expedition to evaluate our competitors in the financial space.

Sun 11:57 AM

We’ve presented to four judges, which is all we’re gonna get I think. Eric takes off to go eat with his friend. I am all alone running the booth. Good thing there are always more Chinese flash cards to do.

Sun 1:05 PM

There are no more judges. It is a desolate wasteland. Why were we placed at the very end of the room.

Sun 1:42 PM

I have discovered that our search algorithm is not deterministic. Why. How is this possible.

Sun 1:53 PM

Nextock is a top 5 finalist. This is not an original idea. This is not a good idea. I hate hackathon judging. (Ed note: Nextock was an app that analyzed news sentiment about a company and predicted their stock price. This idea is as old as machine learning.)

Sun 2:40 PM

Chris Brown Denny and his team just won a prize. They literally spent the whole time memeing. Their submission was a meme. Eric is losing it. He thinks this is no longer a safe space. He’s clutching my blanket ready to break down.

Sun 2:50 PM

We win the HBK prize - beats by Dr Dre. We don’t get the Capital One prize - $200 in Amazon gift cards. Nextock wins that one. I hate them. I would have preferred the reverse. Waiting on 2σ.

Sun 2:52 PM

2σ prize goes to Social Media Stocker. It does exactly what you think: predicts stock prices based on social media sentiment. This is dumb.

Sun 2:56 PM

Third place is PySight. This is a python flask web app wrapper around a object detection and voice to speech API. It does literally nothing interesting. Second place is Cruise. They are the only ones with an interesting project. Brake and turn lights on a longboard. First place is STIC: a selfie stick with a phone and ultrasonic sensor for the blind. Hooks into Microsoft object detection API. HackRice is concluded. This is the shoddiest batch of winners I’ve ever seen in a hackathon. Time to go home.


I’ve been getting flack for such a negative post, so I do want to make a small ammendment: I love hackathons, realy enjoy going to them, and don’t go for the prizes. The prizes here were awarded dissapointingly, but I still had a great time. Overall it was a great event!

See the final product here: YABA

Written on September 24, 2017

Leave a comment!